AI的国情咨文：我们已越过拐点，黑暗工厂即将来临，自动化时间表 | 西蒙·威利森

我很好奇，人类大脑是否有一天根本不再需要，而这正是所有讨论的核心。

And I wonder if the human brain, it s not like maybe someday we don need human brains at all and that s all of the discussion.

但也许下一阶段是，人工智能将帮助我们产生出色的想法。

But maybe the next phase is AI will help us come up with great ideas.

我的意思是，这种情况已经

I mean, that's been

持续了大概好几年了。

the case for probably a couple of years now.

它们已经足够强大，能够进行非常出色的头脑风暴。

They've been strong enough to do really good brainstorming.

我喜欢把它比作这样的情况：当你组织一场小组头脑风暴时，你会预订一间会议室一小时，准备一块白板，召集十几个人。

And I like to compare it to the thing where when you've got a group brainstorming exercise, you book a meeting room for an hour, you've got a whiteboard, You get a dozen people in.

而在那整个头脑风暴过程的前三分之二时间里，老实说，大家只是在反复提出最明显、最基础的想法。

And the first two thirds of that brainstorming session, honestly, it's kind of just everyone going through the most obvious basic ideas.

对吧？

Right?

然后把它们全都写在白板上。

And you get them all out on the whiteboard.

把它们全部列出来。

You get them all up.

当你开始说：好吧，我们来谈谈这些的时候，事情就变得有趣了。

And then things get interesting when you start saying, okay, well, let's talk about these.

让我们开始把它们组合起来。

Let's start combining them.

AI 在创意的前三分之二部分非常擅长。

The AI is so good at that first two thirds of the ideas.

我经常和它一起头脑风暴，只是让它把所有显而易见的想法都吐出来，它能给出二十个点子，而且都挺完整的。

Like, I brainstorm with them all the time where I just get them to spit out all of the obvious stuff, and they'll come up with 20 things, and they'll all be kind of done.

它们就是……它们不会，它们就是不会太有趣。

Like, they're very they won't be they just won't be very interesting.

真正有趣的是，当你再让它给出二十个更多点子时，到了这份列表的末尾，你开始得到一些并不算好、但却能引你走向有趣方向的想法。

What gets interesting is when if you ask them for 20 more, and now they by the sort of end of that list, you're beginning to get things which are not good ideas, but they point you in interesting directions.

还有很多类似的技巧。

There are so many other tricks like this.

比如，你可以让AI把一些奇怪的领域结合起来。

Like, you can tell you can you can tell AI to combine weird fields.

你可以这么说。

You can say, okay.

我想为我的新SaaS平台寻找一些受海洋生物学启发的营销创意，看看会发生什么。

I want ideas for marketing my new SaaS platform inspired by marine biology, and you see what happens.

大部分内容可能会是垃圾，但可能会有一个火花让你找到好点子。

And most of it will be complete junk, but there might be a spark that gets you to the good idea.

所以我非常喜欢把AI当作这方面的头脑风暴伙伴。

So I love them as as brainstorming companions on that front.

这让我想起我和大卫·普拉塞克的一次聊天。

That reminds me of a chat I had with David Plasek.

他是位命名专家。

He's a expert naming person.

他帮助公司为产品起名。

He helps companies come up with names for products.

他公司的一项工作是组建三个团队来头脑风暴产品名称。

And one of the things that he does at his company is he creates three teams to come to brainstorm names.

比如，假设他们为一个帆板运动产品命名。

One team so for example, let's say a windsurf was a product they named.

第一组的想法是，好吧，这是一个AI开发工具。

So the first team is, okay, this is an AI IDE thing.

这正是它的本质。

That's that's exactly what it is.

第二组的想法是，好吧，这是一艘船。

Second team is okay, this is a this is a boat.

你在为一艘船命名，这里有若干限制条件。

You're naming a boat and here's constraints.

而这里，这是一艘宇宙飞船。

Then here, this is a spaceship.

所以从这个角度来命名。

So name it from that perspective.

他发现，最好的名字往往来自这些不同的视角，即用不同的隐喻表达相似的好处。

And he finds the best names come from those other directions where it's a different metaphor with the same sort of benefits.

好的。

Okay.

所以我的理解是，这很不错。

So what I'm hearing here is this is good.

目前这对人类来说是好事，因为我们仍然有机会参与到这个过程中。

This is good for humans right now, that there's still opportunity for us to contribute to the process.

实际上，我想为软件工程师辩护一下，因为一方面，这些工具确实能写代码。

And actually, I want to stand in defense of software engineers for a bit because on the one hand, these things can write code.

这曾经是我们的专长。

That used to be our thing.

对吧？

Right?

我发现，很好地使用编码代理，需要动用我作为软件工程师二十五年来的全部经验，这让我精神上非常疲惫。

I'm finding that using coding agents well is taking every inch of my twenty five years of experience as a software engineer, and it is mentally exhausting.

这种事情现在人们谈论得越来越多了。

Like, this is something which people are talking a lot more about now.

我可以同时启动四个代理，让它们分别处理四个不同的问题。

I can fire up, like, four agents in parallel and have them work on four different problems.

到了上午十一点，我就已经精疲力尽了。

And by, like, 11AM, I am wiped out for the day.

因为人类的认知能力是有极限的，即使你不去审查它们做的每一件事，单是同时在脑海中保持多少信息，就已经很容易超出负荷了。

Like, I have because there is a limit on human cognition in how much even if you're not reviewing everything they're doing, just how much you can hold in your head at one time, and it's very easy to pop that stack at the moment.

我们需要学习一种新的个人技能，那就是找到自己的新极限。

Like, there's a sort of personal skill that we have to learn, which is finding our new limits.

我们该如何负责任地使用时间，避免耗尽自己，这才是关键。

Like, what is what is a responsible way for us to you to to not burn out and for us to to use the time that we have.

我接触过很多人，他们因为失眠而困扰，因为他们觉得：我的编码代理本可以替我完成更多工作。

And I I've I've talked to a lot of people who are losing sleep because they're like, my coding agents could my agents could be doing work for me.

我只是打算多熬半小时，再启动一堆任务，让它们在凌晨四点醒来工作。

I'm just gonna stay up an extra half hour and and set off a bunch of extra things, and they're waking up in four in the morning.

这显然是不可持续的。

That's obviously unsustainable.

我希望这只是一个暂时的新鲜现象。

I hope that that's a novelty thing.

这些代理工具真正变强大，也不过是过去四五个月的事。

The agents only really got good in the past sort of four to five months.

我们都在摸索，这到底意味着什么，能让我们做到什么。

We're all learning what that looks like and what that lets us do.

但这确实令人担忧。

But it's it's it's concerning.

我们在使用某些工具时，带有一种赌博和上瘾的成分。

There's an element of sort of gambling and addiction to to how we're using some of these tools.

但为了为软件工程师辩护，我从这些工具中获得了极佳的成果，因为它们确实是现有技能和经验的放大器。

But to stand in defense of software engineers, I get great results out of these things because they are amplifiers of existing skills and experience.

我有二十五年AI之前的丰富经验，现在可以通过与代理进行高层次对话来放大这些经验。

And I have twenty five years of existing, like, pre AI experience, which I can now amplify because I can talk to the agent at a very high level.

我可以使用那些我多年掌握的复杂工程语言，而代理似乎也能理解这些语言。

I can use very I can use sophisticated engineering, like, language that I've mastered over the years, which they appear to know as well.

我们可以非常高效地协作。

And we can collaborate incredibly effectively.

这意味着我可以审视一个问题，然后说：这个问题只需要一句提示，我知道代理能找出并修复这个bug，而另一个问题则不确定到底有多大。

That means I can look at a problem, and I can say this problem is a one sentence prompt, and I know it'll find that bug and fix that bug, as opposed to this other problem, which is who knows how how big a problem.

但另一方面，我有二十五年对开发耗时的判断经验，而这些经验现在完全失效了。

There is a flip side to this, which is that I've got twenty five years of experience in how long it takes to build something, and that's all completely gone.

因为现在我可以看一个问题，然后说：好吧，这个大概需要两周时间。

Like, that doesn't work anymore, because I can look at a problem and say, okay, well, this is gonna take two weeks.

不值得做。

It's not worth it.

这就像，是的。

And that's like, yeah.

但也许这只需要二十分钟，因为原本需要两周的原因是那些繁琐的编码工作，现在都被AI替我们完成了。

But maybe it's gonna take twenty minutes because the reason it would've taken two weeks was all of the the sort of crafty coding things that the AI is now covering for us.

我发现这非常有趣且具有挑战性。

And that I've been finding really interesting and challenging.

我不断给AI分配一些我认为它无法完成的任务。

I I constantly throw tasks at AI that I don't think it'll be able to do.

因为偶尔它真的能完成。

Because every now and then it does it.

当它没能完成时，你就学到了，对吧？你知道了，Opus 4.6 仍然无法完成这个特定任务。

And when it doesn't do it, you learn, right, you learn, okay, Opus 4.6 still can't do this particular thing.

但当它完成了某些以前模型做不到的事情时，这实际上就是前沿的AI研究。

But when it does do something, especially something the previous models couldn't do, that's actually cutting edge AI research.

你可能是世界上第一个发现AI现在能做某件事的人，因为你曾经发现它做不到，于是你一直保留着这些有趣任务的清单。

You can be the first person in the world to spot that AI can now do x just because you were the person you you found it couldn't do it, you've you've been keeping that sort of backlog of of interesting tasks for it.

这个讨论方向太有意思了。

This is such an interesting line of discussion.

你在这里描述的正是这样一种观点：比如所谓的10倍工程师会变得更有价值，因为你能更有效地使用这些工具。

This idea that, let's say, 10x engineers, to use that phrase, are gonna be more valuable is what you're describing here because you can work with these tools much more effectively.

你对初级工程师怎么看？

What do you think of junior engineers?

那边正在发生什么？

Just like what's happening there?

他们的未来会怎样？

What's their future?

有意思的是，Thoughtworks这家大型IT咨询公司大约一个月前举办了一次外出研讨会，他们邀请了来自不同公司的多位工程副总裁来讨论这个问题。

So there's an interest so Thoughtworks, the big, like, IT consultancy did a off-site a few, about a month ago, and they produced they got a whole bunch of engineering VPs in from different companies to talk about this stuff.

他们提出的一个有趣观点是，这类工具对资深工程师特别有帮助。

And one of the interesting theories they came up with is they think this stuff is really good for experienced engineers.

它能放大他们的技能。

Like, it amplifies their skills.

这很棒。

That's great.

这对新工程师非常有帮助，因为它解决了许多入职过程中的问题。

It's really good for new engineers because it solves so many of those onboarding problems.

比如，如果你去问问 Cloudflare 和 Shopify，他们都表示，2025 年一整年计划招聘一千名实习生，因为过去实习生需要一个月才能开始做有用的工作。

Like, if you talk to, Cloudflare and Shopify, both said they were hiring a thousand interns over the course of 2025 because the intern onboarding costs it used to be takes a month before your intern can do anything useful.

但现在，他们一周内就能开始做有用的事情了，因为 AI 助手能帮助他们更快上手。

Now they're doing something useful within like a week because the the AI assistant helps them get up and running faster.

问题在于中间那批人。

The problem is the people in the middle.

比如，如果你是中阶工程师，还没达到资深工程师的水平，但也不算新人了，那正是 ThoughtWorks 目前最可能陷入困境的群体。

Like, if you're mid career, if you haven't made it to sort of super senior engineer yet, but you're not sort of new either, that's the that's the group which ThoughtWorks which ThoughtWorks Resolve were probably in the most trouble right now.

这是一个悬而未决的问题，因为他们缺乏足够的经验来借助这些工具提升效率，也不像新人那样能获得那么多明显的助力。

Like, that's the open question because they don't have that expertise to to to to amplify and and use with these tools, and it's not as benefit like, they've got all of the the boosts that the beginners were getting they've got already.

所以，这对我来说是一个当前有趣的开放性问题。

So that's an interesting open question right now for me.

问题更多集中在中阶工程师身上，而不是新人或高级工程师。

It's it's more the the the sort of mid mid level as opposed to the beginners or the the advanced people.

人工智能出现在这么多事物的中间，这真是太有趣了。

It's so interesting how AI is coming at the middle of so many things.

它出现在产品开发流程的中间。

It's coming at the middle of the product development process.

它出现在资历的中间。

It's coming at the middle of seniority.

可能还有其他例子。

There's probably other examples.

我猜这对所有职能都适用，比如产品经理和设计师，尤其是新入行的产品经理和设计师，因为你们所说的‘AI原生’本质上就是能更快上手。

And I'm guessing this is true for all functions like PMs, designers too, just new PMs, designers, maybe because being AI native basically is what you're describing and ramping up much more quickly.

我想既然我们谈到这个话题，很多听众其实就属于那些处于中间阶段的人。

I guess while we're on this topic, say you are a lot of listeners here, just like those people in the middle.

你对他们有什么建议，可以帮助他们避免沦为永久的底层群体？

What would your advice be to them to help them avoid becoming a part of the permanent underclass?

你这是给我施加了很大的责任啊。

That's a big responsibility you're putting on me there.

我认为前进的方向是拥抱这些技术，弄清楚如何利用它们让我变得更好？

I think I think the way forward is to lean into this stuff and figure out how do how do I help this make me better?

对吧？

Right?

很多人担心技能退化。

Like, a lot of people worry about skill atrophy.

你知道，如果AI替你做了，你就什么也学不到。

You know, if the AI is doing it for you, you're not learning anything.

我认为如果你担心这个，就应该对它保持警惕。

I think if you're worried about that, you push back at it.

你必须谨慎地思考如何应用这项技术，想想：我得到了一个能回答任何问题、而且经常答对的东西，虽然它并不总是对的。

Like, you have to be mindful about how you're applying the technology and think, okay, I've been given this thing that can answer any question and often gets it right, doesn't always get it gets it right.

我该如何利用它来放大自己的能力，学习新东西，承担更雄心勃勃的项目？

How can I use this to amplify my own skills, to to learn new things, to take on much more ambitious projects?

我最近很喜欢的一点是，作为一名软件工程师，我发现自己的抱负水平直线上升。

Something I've been enjoying I think the thing I've enjoyed most about this as a software engineer is that my level of ambition has shot right up.

因为我以前从不使用AppleScript，因为AppleScript是一整套需要学习的编程语言。

Because now I used to, like never I never used AppleScript because AppleScript is a whole programming language you have to learn.

但过去两年半我一直在用AppleScript，因为ChatGPT懂AppleScript，我不用自己学，现在我能自动化我的Mac上的各种任务。

And I've been using AppleScript for, like, two and a half years now because ChatGPT knows AppleScript, and I don't have to and so now I can automate things on my Mac.

这很棒。

And that's great.

你知道吗？

You know?

以前，光是想到要花两三个月才能掌握基本的AppleScript，就足以让我根本不想去用它。

And previously, the fact that it would've taken me, like, two or three months to learn basic AppleScript was enough for me never to use it.

而现在，我用了这么多新技术，就是因为原本两到三个月的初始学习曲线被大大缩短了。

And now I've got all of these technologies that I'm using because that two to three month initial learning curve has been shaved right down.

我觉得这适用于所有其他事情。

I think that applies to everything else.

比如，我的厨艺进步了很多。

Like, I'm getting much better at cooking.

我一直在用Claude，结果发现它是个出色的厨师，这听起来很奇怪，因为它根本没有味蕾。

I've been using Claude, it turns out, excellent chef, which doesn't make sense because it can't it doesn't have taste buds.

但它能提供全球牛油果酱食谱的平均版本，结果发现这确实是不错的牛油果酱。

But it does it can give you the global average of the world's guacamole recipes, which turns out is good guacamole.

所以这真的很有趣，就像尝试把这些技术应用到自我提升上。

So that's been really interesting, like trying to apply this stuff just to for sort of self improvement.

我认为这是一种非常有用的技能。

I think that's a really useful skill to have.

因为老实说，现在一切都变得太快了。

Because honestly, everything is changing so fast right now.

唯一的通用技能就是能够适应变化。

The only universal skill is being able to roll with the changes.

对吧？

Right?

这是我们所有人都需要的能力。

That's the thing that we all need.

奇怪的是，在这些关于如何与人工智能相处得更好的对话中，出现最多的词是‘自主性’。

Weirdly, the term that comes up most in these conversations about how you can be great with AI is agency.

对吧？

Right?

人类拥有自主性，我们用这种自主性来决定要应对哪些问题以及往哪个方向走。

People, human beings have agency, and we use that agency to decide what problems to take on and where to go.

我认为人工智能根本没有任何自主性。

I think agents have no agency at all.

我认为，人工智能永远不可能拥有的东西就是自主性，因为它没有人类的动机。

Like, I would argue that the one thing AI can never have is agency because it doesn't have human motivations.

当然，你可以告诉它要赚更多钱之类的，但它永远无法自己决定接下来什么行动才是合理的。

Like, sure, you can tell it make more money or whatever, but it's never going to be able to decide on its like, what makes sense for it to act on next.

所以我认为，关键在于投资自己的自主性，思考如何利用这项技术来提升自己的能力，去做新的事情。

So I'd say that's the thing is to invest in your own agency and invest in how do I use this technology to get better at what I do and to do new things.

而且，正如你所说，要有雄心。

And also to your point, be ambitious.

志存高远。

Think big.

对。

Yeah.

昨天刚发布了一段对詹森的采访，有人问他关于裁员的问题。

There's an interview with Jensen that just came out yesterday where people asked him about layoffs.

现在到处都在裁员。

There's all these layoffs happening.

AI真的在夺走工作吗？

Is AI actually taking jobs?

他说，很多公司之所以裁员，是因为他们缺乏足够的创造力和雄心，不知道如何利用这些资源。

And he's like, the reason a lot of these companies are not, are letting people go is they don't have enough creativity or ambition for what they can do with all of these resources there.

因为他们并没有裁员。

Because they're not letting people go.

他们有太多想做的事情。

They have so much they want to do.

你知道，说起来容易做起来难，而且并不总是如此，但我认为这是一种很有意思的处理方式。

You know, obviously easier said than done and it's not always the case, but I think that's an interesting way of approaching it.

既然我们现在拥有这种能力，人们几乎低估了自己能用它做什么，没有充分去利用它。

Now that we have this power, people almost underestimate what they can do with it and don't fully lean into it.

所以我非常喜欢这种建议：试着更有雄心一点。

So I love this advice of just try to be a little more ambitious.

去尝试那些你觉得不可能的事情，看看它们可能其实并非不可能。

Try the stuff that you think is impossible and see might be actually possible.

我今年的新年决心恰恰相反。

My New Year's resolution this year was the opposite.

过去每一年，我总是告诉自己：今年我要更专注。

Every previous year, I've always told myself this year, I'm gonna focus more.

我要少接一些事情。

I'm gonna take on less things.

但今年，我的目标是多接一些事，更有雄心。

This year, my ambition was take on more stuff and be more ambitious.

我们有这些工具。

Like, we've got these tools.

把所有东西都拿过来。

Bring it all in.

让我们试试做所有事情。

Let's try and do everything.

我不知道这是否是个好的新年决心，但我就选了这个。

I don't know if that was a good New Year's resolution, but that's what I went with.

那么到目前为止进展如何？

So how's it going so far?

你对

How do you feel about

这个决定感觉如何？

this decision?

很有趣。

It's fun.

我玩得很开心。

I'm enjoying myself.

我觉得我可能会坚持到年底，但那时候我会想，哇。

I I think I'll probably get to the end of the year, but I'll be like, wow.

我本该重点关注的最重要的事情都没有完成。

The thing the most important things that I should have been focusing on did not get done.

但当我有志于做这些事时，情况就是这样。

But that's that's the case when it is my ambition to do them.

所以，你知道的。

So, you know.

这是一种聚散交织的情况。

It's a a converge diverge sort of situation.

你知道的。

You know?

明年可以重新聚焦。

Next year could be refocused.

当然。

Absolutely.

对。

Yeah.

天哪。

Oh, man.

有点吧。

Kind of

就是那种情况。

along those lines.

我想回到你提到的这一点，你说你工作更努力了，但一整天早早就精疲力尽了。

I want to come back to this point you made about how you're working harder and you're like fried early in the day.

这真是个有趣，我不知道怎么说，几乎是种矛盾。

This is such an interesting, I don't know, contradiction almost.

人们都知道，人工智能本应让我们更高效。

People, you know, AI is supposed to make us more productive.

它本应给我们更多休息时间。

It's supposed to give us more time off.

它本应让我们闲坐着看奈飞，享受世界上所有的财富与生产力。

It's supposed to let us sit around and watch Netflix and do all the create wealth and productivity in the world.

感觉那些最沉迷于AI的人，反而比以往工作得更努力。

It feels like the people that are most AI pilled are working harder than they've ever worked.

有一种焦虑被描述为：我的代理没有运行。

There's this anxiety described of my agents aren't running.

我得时刻盯着它们。

I gotta stay on top of them.

你认为这里发生了什么？

What do you think is going on there?

这就像你所说的，可能只是一种暂时的新鲜感，然后我们会说，好吧。

Is this just like you said, maybe it's like a temporary novelty thing, and then we'll be like, alright.

我不需要这么有生产力。

I don't need to be this productive.

还有别的吗？

Is there anything else there?

我希望这真的只是一时的新鲜感。

I think I I really hope it's a novelty thing.

我确实有了更多时间，但我却感到精疲力尽。

And I am actually getting much more I'm getting more time, but I'm I'm exhausted.

你的大脑累坏了。

Like, your brain is exhausted.

我的大脑累坏了。

Like, my brain is exhausted.

我有更多时间去做其他事情，我也确实去做了，这很棒。

I've got I've got more time to go and do things, and I do things, and it's great.

但这种高强度工作带来的疲惫感，真的让我很意外。

But it's it is that the exhaustion from that sort of intensity of work has been a really big surprise for me.

尤其是从十一月开始，随着这一切逐渐升温，我一直在观察这种现象。

Like, that that's been been some something which I've I've I've I've been observing, especially since November, like, as as all of this stuff stuff started ramping up.

而且，是的，我认为这个问题的核心始终是来自他人的期望。

And, yeah, I think that's the concern there comes down it's always expectations from other people.

你知道的？

You know?

如果你在一家期望你完成五倍工作量的公司，那肯定会让人精疲力尽。

If you work for a company that's that's expecting you to get five times more done, that's gonna be exhausting.

也许我们以后会看到。

And and maybe we'll see.

我认为那些管理良好的好公司正在关注这个问题。

And I think the good companies with good management are paying attention to this.

它们不希望为了短期利益而让最优秀的员工 burn out，最终却因此失去人才。

They don't want to burn out their best employees for the sort of for the short term gain, but but lose people over it.

但，是的，这确实是一个巨大的张力。

But, yeah, it's it's it's a big tension.

我认为，我们这些身处人工智能热潮前沿的人，是最早感受到这一点的。

I think we're we're those of us on the sort of leading edge of the AI boom are feeling it first.

我猜这对其他人也会发生。

I imagine it's gonna come for everyone else as well.

不过，我们还没提到的另一个因素是，你已经提过几次了。

The other element of this, though, that we haven't mentioned is and you've mentioned a couple times.

这实际上非常有趣。

It's actually really fun.

这里的动力并不是

The drive here is not

我自己也做了很多。

I have to myself so much.

当然。

Absolutely.

这太有趣了。

It's so fun.

我的很多朋友都在谈论他们有一堆积压的副项目。

It's a lot of my friends have been talking about how they have this backlog of side projects.

对吧？

Right?

过去十到十五年里，他们一直积压着一些从未完成的项目和觉得会很酷的想法。

For the last ten, fifteen years, they've got projects they never quite finished and ideas they thought would be cool.

其中一些人说：好吧，我现在全都做完了。

And some of them are like, well, I've done them all now.

就在过去几个月，我一一处理了。

Like, last couple of months, I just went through.

每天晚上，我都想：来吧，把这个项目做完。

And every evening, I'm like, let's take that project and finish it.

还有这个，那个，还有那个，以及那个。

And that one and that one and that one and that one.

当一切结束时，他们几乎感到一种失落感：好吧，就这样了。

They almost feel a sort of sense of loss at the end where they're like, well, okay.

我的待办清单没了。

My backlog's gone.

现在，我该去构建什么呢？

Now now what am I gonna build?

是的。

Yeah.

这又回到了那个工厂的话题。

It comes back to that factory.

前几天我跟Linear的创始人聊了聊，谈到‘工厂’这个概念，我们觉得，‘工厂’听起来不像能创造出优秀产品的地方。

I was talking to the founder of Linear the other day and this idea of the factory, and we're just like, like, factory doesn't sound like a place that'll create amazing products.

感觉好像，这种地方能诞生出美丽而创新的东西的可能性有多大呢？

It feels like, you know, like, what are the chances that'll create something beautiful and innovative?

所以，要么这个词用错了，要么它只会导致糟糕的结果，大概率是这样。

So either that's the wrong word or it's just this will lead to bad stuff, probably.

我觉得‘手工制作’这个词更好，手工制作的软件，我认为会更受重视。

I feel like the word artisanal does like like, artisanal's a handcrafted software, I think, is gonna be valued more.

我在自己的工作中注意到，有时候我会对某个软件、某个Python库产生一个想法，然后一小时内就能完成，还能附上文档和测试。

Something I've noticed in my own work is sometimes I have an idea for a piece of software, a Python library or whatever, and I can knock it out in, an hour and get to a point where it's got documentation and tests and all of those things.

而且它看起来就像我之前花了好几个星期做的那种软件。

And it looks like the kind of software the previous I just spent several weeks on.

我可以把它上传到 GitHub 上，一切都没问题。

And I can stick it up on GitHub and everything.

但我并不相信它。

And yet, I don't believe in it.

我不相信它的原因是，我不得不匆忙完成所有这些工作。

And the reason I don't believe in it is that I I got to rush through all of those things.

我觉得质量可能还不错，但我没有花足够的时间去验证这种质量。

I think the quality is probably good, but I haven't spent enough time with it to to feel confident in that quality.

最重要的是，我还没用过它。

Most importantly, I haven't used it yet.

事实上，当我使用别人的软件时，我最在意的是希望他们已经用了好几个月。

Like, it turns out when I'm using somebody else's software, the thing I care most about is I want them to have used it for for months.

对吧？

Right?

我希望其他人能把这些软件真正用起来。

I want other people to have put that software into practice.

我有一些非常酷的软件，是我自己写的，但我从来没用过。

So I've got some very cool software that I built that I've never used.

说实话，写它比真正去用它还要快。

Like, it was so it was quicker to build it than to actually try and use it.

所以，我一直以来应对这种情况的方式，就是给它打上alpha标签。

And so the way I've been dealing with that, I've always put alpha on it.

比如，如果你看到我的软件标注为alpha，那很可能意味着我还没真正用过它——对我大部分项目来说都是这样，这简直是个取巧的招数，你知道的，动不动就标个alpha。

Like, if you see my software and it says it's an alpha, that probably means I haven't actually used it yet for most of my projects, which is a bit of a cheat code, you know, alpha alpha this.

但这难道不有趣吗？

But isn't that interesting?

以前，如果你看到一款软件有高质量的测试和详尽的文档，那就意味着它很可靠。

Like like like, it used to be if you looked at software and it had high quality tests and documentation and everything, it meant it was good.

但现在，这个信号已经消失了。

And now that signal is gone.

这简直像是我们需要一个工作证明，而不是

It's almost like we need a proof of work for this versus

使用证明。

the Proof one of usage.

证明，是的。

Proof Yes.

正是如此。

Of Exactly.

天哪。

Oh man.

说到手工编写的代码，我不知道你是否知道这一点，这太有趣了。

On this note of handcrafted code, I don't know if you know this, this is so interesting.

数据标注公司正在购买旧的 GitHub 手写代码仓库来训练他们的模型，并且为这种手工编写的人类代码支付了大量资金。

Data labeling companies are buying old GitHub repos of handwritten code to train their models on, and they're paying a lot of money for like artisanal human written code.

哦，这太有趣了。

Oh, that's fascinating.

那是二战前的金属，你可以从旧船残骸中打捞出来，那是第一次核爆炸之前的金属。

That's the the pre World War two the the metal that you can dig up from old shipwrecks, which is before the nuclear the first nuclear explosions.

所以，这种金属没有受到辐射的污染。

And so it's it's not got, like, the the the the radiation bakes into the metal.

就是整个这么一回事。

It's that whole thing.

哇。

Wow.

这太有趣了。

That's fascinating.

是的。

Yeah.

所以他们在寻找2022年之前的代码，我想，也就是ChatGPT刚出现的时候之前的代码。

So they're looking for code pre 2022, I think, whenever ChatGPT kind of emerged.

哇。

Wow.

是的

Yeah.

所以如果你有一些，你就能赚大钱。

So if you've got some, you can make a you can make a fortune.

我保证。

I promise.

我开源了我所有的东西，所以它们已经公开了。

I open source all my stuff, so it's already out there.

它就在训练数据中。

It's it's in the training

它已经被用来训练模型，早就被用过了。

it's it's been used to train the models slurp stuff already.

没错。

Yep.

哦

Oh,

天啊。

man.

好吧。

Okay.

我问你一个问题。

Let me ask you this question.

我只是对这个预测感到好奇。

I'm just curious about this prediction.

我知道你并不像那种喜欢做预测的人，尽管你确实经常做出预测，而且你似乎总是对的。

I know you're not like a prediction person, although you do make predictions and you seem to be right often.

你认为世界上50%的工程师什么时候会由AI编写他们100%的代码？

When do you think 50% of engineers in the world will be AI will be writing a 100% of their code?

你觉得我们离这个目标还有多近？

How close to that do you think we are?

那我把它修正为95%的代码。

So I'm gonna refact that to 95% of their code.

我不会

I don't

认为是的。

think Yeah.

我们会谈到这一点的。

We'll get to that.

但没错。

But yes.

很难说全球范围的情况，部分原因是不同文化之间存在差异。

It's very difficult to say worldwide because partly because to their cult there are cultural differences.

我花了太多时间在 Hacker News 上，我注意到一个现象：在太平洋时间午夜开始的对话，会一直持续到早上8点。

I had spent way too much time on hacker news, and something I've noticed about hacker news is a conversation that starts at midnight Pacific time and goes until 8AM.

语气非常不同，因为那是欧洲人。

Very different tone because it's the Europeans.

对吧？

Right?

你会看到，欧洲人比美国人更怀疑人工智能。

You'll get the year and the Europeans are a lot more AI skeptic than the Americans are generally.

所以我认为不同国家将会形成各自不同的文化氛围。

So I think different countries are gonna have different sort of different cultures around us.

同时，我认为今年已经毫无疑问地表明，这些东西能生成优质的代码。

At the same time, I think it's become undeniable this year that this stuff produces good code.

过去你可以说，我不用这些东西，因为代码质量差。

Like it used to be that you could say, I don't use this stuff because the code is bad.

那曾经是一个合理的立场。

And that was a a justifiable position.

但现在这已经不再合理了。

That's not justifiable anymore.

代码现在很好了，至少按照我对好代码的定义来说是这样。

The code is now It's good code for for the my for my definition of good code, at least.

所以我们说，到今年年底，可能有50%的工程师，他们的大部分代码都可能由AI生成。

So so we're saying 50% of engineers let's say 50% of engineers majority of their code, it could happen by the end of this year.

有可能。

It could.

因为现在的技术已经足够好了。

Because the the the the technology is good enough now.

我觉得现在的挑战是让人们学会如何使用这些工具，这很难，因为一提到使用这些工具，大家都觉得肯定很简单。

And I feel like the the challenge now is getting people to learn how to use this stuff, which is difficult because using the stuff, everyone's like, oh, it must be easy.

不就是个聊天机器人嘛。

It's just a chatbot.

其实并不简单。

It's not easy.

在人工智能领域，一个很大的误解就是认为有效使用这些工具很容易。

Like, that's one of the great misconceptions in AI is that using these tools effectively is is is easy.

这需要大量的练习，需要不断尝试那些没成功的方法，也要尝试那些成功的方法。

It takes a lot of practice, and it takes a lot of trying things that didn't work and trying things that did work.

但是的，我预计到今年年底，工程师说他们几乎所有的代码都是由AI编写的，将不再罕见。

But, yeah, I I I expect by the end of this year, it will not be uncommon to have an engineer say that almost all of their code is written by AI.

我之前也有过类似的想法。

That was the same rough idea I had.

这难道不疯狂吗？

And how crazy is that?

变化速度真快，太惊人了。

How quickly It's wild.

这份工作已经发生了变化，可能性也不同了。

This job has changed and what is possible.

我认为人们常常低估了事物变化的速度，这是一个很好的例子。

I think people, this is a good example of people underestimate how quickly things can change.

一年前或两年前，达里奥还在预测这种情况，说100%的代码都将由AI编写，当时我们还觉得这不可能。

We would not have, I think Dario was predicting this a year or two ago, just so, 100% of code's going be written by AI, and we're just like, We

我们很乐意向蒂姆致敬。

will love to Tim.

是的。

Yeah.

对吧？

Right?

没错。

Exactly.

你在说什么啊？

Like, what are you talking about?

太差了。

So bad.

写代码水平太差了，而且这种变化可能会出现在人们没预料到的其他工作中，这既令人害怕，又有趣且令人兴奋。

So bad at writing code, and and this might come for other jobs that people don't see coming, which is scary and interesting and exciting.

说实话，我完全不是AI末日论者。

It's honest the the I'm I'm not an AI doomer in the slightest.

但从经济角度看，确实让我感到不安。

The economics of it do make me nervous.

我们的白领知识型工作，真的会在未来几年内被消灭十分之一吗？

Like, it are we really going to wipe out, like, a tenth of white collar knowledge work jobs in the next few years?

我真的希望不会，因为我不知道经济如何适应这种情况。

I really hope not because I don't know how the economy adapts to that.

你知道的。

You know?

所以，是的，这很复杂。

So, yeah, that's complicated.

是的。

Yeah.

我其实正在做一份即将发布的报告。

I'm actually I'm doing a report that's coming out.

它会在本集播出前发布，分析科技行业的就业市场。

It'll come out ahead of this episode, looking at the job market in tech.

令人惊讶的是，仅在科技公司，我们目前的工程岗位和产品经理岗位空缺数量达到了历史最高水平。

And surprisingly, just at tech companies, we're at the highest number of open engineering roles, open PM roles.

有意思。

Interesting.

仅有的例外是新冠疫情期间那段疯狂的峰值时期。

Except for during the crazy peak during COVID.

所以现在的情况有点往那个水平回升了。

So it's kind of like coming back to that.

基本上来说，全球科技公司的工程师和产品经理空缺职位数，是这三年半左右以来的最高值。

Basically, it's the highest number of open roles in three and a half ish years for engineers and PMs at tech companies globally.

这太有意思了。

That's very interesting.

这挺有意思的，不是吗？

It's funny, isn't it?

因为媒体上全是这些博眼球的裁员新闻。

Because you get all of these headline grabbing like Layoffs.

对。

Yeah.

最近是不是Block那家公司裁了4000人来着？

Was it was it Block that laid off 4,000 people recently?

是的。

Yeah.

这里始终存在的问题是，这其中有多少是人工智能的影响，有多少是疫情期间过度招聘和后续调整造成的。

The the the the the question there is always how much of that is AI and how much of it is over hiring during COVID and re corrections and all that kind of thing.

要分清楚总是非常困难。

It's always very difficult to tell.

所以，一方面，空缺职位的数量可能是一个更好的信号。

So that the the number of open jobs, on the one hand, maybe that's a better signal.

但另一方面，招聘市场已经被这一切彻底搅乱了。

But on the other hand, the recruitment market has been driven completely crazy by all of this stuff.

对吧？

Right?

所有的职位广告都是由人工智能撰写的。

Like, all of the job ads are written by AI.

简历也是人工智能生成的。

The the the the resume's AI.

招聘领域的人表示，现在筛选和雇佣人才的难度前所未有，而求职者则说他们投了200份简历，却连一个回音都没有。

People people in recruitment are saying that this is it's never been this hard to filter through and hire people, and people who are hiring jobs say they they applied to 200 things and got nobody hearing back.

所以这很难。

So it's hard.

对吧？

Right?

关于这个问题的宏观经济指标都是滞后的，迟早我们会得到更可靠的数字，来了解实际影响究竟如何。

The the the the macroeconomic indicators for this stuff are are lagging, and at some point, we should start getting more confident numbers about what the impact actually is.

是的。

Yeah.

有趣的是，招聘岗位的空缺数量也接近历史最高水平。

Interestingly, the number of recruiter open roles is also approaching like record numbers.

太荒谬了。

Hilarious.

这其实是一个反映招聘需求的有趣先行指标。

Which is an interesting leading indicator of demand for hiring.

尽管有裁员，但仍有一些有趣的趋势。

So there's interesting trends in spite of the layoffs.

所以，真是个疯狂的世界啊。

So yeah, what a wild world.

你提到过你正在写的一本书。

So you've mentioned this book you're working on.

这说的是智能体工程模式，对吧？

This is the agentic engineering pattern stuff, right?

是的。

Yes.

好的，太好了。

Okay, cool.

我想聊聊这个。

So I want to talk about this.

你指出，人们以为用AI做开发很容易，觉得AI会替我们做所有事情，那我们整天干什么呢？

So you pointed out, people think it's easy to build with AI, it's like, was going do all these things for us, what are going do all day?

你说得对，实际上并不是这样。

To your point, it's actually not.

要做好这件事，你需要很多非常具体的技能，而你正在你的博客上把这些技能整理出来，我们会引用它。

There's a lot of very specific skills you need to do this well and you're putting them together on your blog, we'll point to it.

我想详细讲讲其中几个，帮助大家更好地完成这件事。

I want to talk through a few of them, to help people do this better.

其中一个就是，现在写代码的成本已经很低了。

So one is this idea of just writing code is cheap now.

你之前稍微提到了这一点。

You touched on this a bit.

也许你可以分享一下，为什么这一点如此重要，值得我们时刻牢记。

Maybe just share why this is such an important thing to know and keep in mind.

我认为这是整个过程中最令人震惊的一点。

So I think this is the single biggest shock in all of this.

我们必须重新思考如何构建、如何作为软件工程师工作，原因在于，过去耗时最长的那部分工作，现在所需时间大大减少了。

The reason that we have to rethink how we build, how we work as software engineers, is that the thing that used to take the time takes way less time.

事实上，程序员从来不会每天花90%的时间在键盘上敲代码。

Like, it's it's never been the case that programmers spend 90% of their day typing code into a computer.

一直以来，围绕编码还有大量其他工作。

There's always there's so much additional work around that.

但过去人们总强调，千万别打断程序员的工作。

But it still used to be like, people talk about how important it is not to interrupt your coders.

对吧？

Right?

你的程序员需要连续两到四个小时不受打扰的时间，才能建立起思维模型并高效写出代码。

Your coders need to have, like, solid two to four hour of uninterrupted work so they can spin up their mental model and and churn out the code.

但这种情况已经完全改变了。

It's some that that's changed completely.

就拿我来说，我编程时每隔一会儿只需要花两分钟向我的助手请教下一步该做什么，然后就能继续做其他事，再回来继续。

Like, I my my programming work, I need two minutes every now and then to prompt my agent about what to do next, and then I can do the other stuff, and I can go back.

我现在比以前更容易被打断了。

I'm much more interrupt ible than I used to be.

但确实如此。

But yeah.

过去耗时的事情，现在所花的时间少了很多。

So the thing that used to take the time is now the thing that takes way, way less time.

这对我们在其他方面所做的一切意味着什么？

What does that mean for everything else that we do?

这不仅仅影响程序员。

And that doesn't just affect programmers.

它影响着整个软件开发团队及其周边的团队。

It affects entire, like, teams of teams around around software development.

但作为个体程序员，你必须开始思考：好吧。

But as an individual programmer, you have to start thinking, okay.

我现在能在写100行代码的时间内写出10000行代码。

I can churn out 10,000 lines of code now in the time that it's saying it take me to write a 100.

我该如何让这些代码变得优质？

How do I make that code good?

对吧？

Right?

我如何确保自己不是在不断产出一堆垃圾，最终积累成拖慢我的技术债务？

How do I make sure that I'm not just churning out total slop that that adds up to technical debt that slows me down?

我如何利用代码现在变得廉价这一事实，来产出更好的代码？

How do I take the fact that code is now cheap and use that to produce better code?

因为我不只是想要廉价的代码。

Because I don't don't just want cheap code.

我想写出真正优秀的代码，能够实现我需要的功能，便于未来扩展，并具备所有那些有用且适用于生产环境的代码特征。

I want really good code that does what I need it to do, like an extend in the future, that's got all of those those characteristics of of of of code that that's that's useful and and can be used in production.

你之前提到的一个观点，我认为在这方面非常重要，那就是当你启动一个项目时，可以同时启动三个不同版本，这有助于你确定方向。

The point you made earlier, I think, is a really important one along these lines, which is when you start a project, you fire off three different versions of it, and that helps you pick a direction.

而这一切之所以可能，是因为现在的代码成本已经非常低了。

And that's only possible because code is so cheap now.

对吧？

Right?

是的。

Right.

我认为原型制作几乎是免费的。

Prototyping is almost free, I think.

这对我影响很大，因为在我整个职业生涯中，我的优势一直是原型制作。

And that really impacts me because throughout my entire career, my superpower has been prototyping.

我非常擅长快速做出可用的原型。

Like, I am very I've been very quick at knocking out working prototypes of things.

我就是那种能走进会议说：看。

I'm the person who can show up at a meeting and say, look.

这是它可能的工作方式。

Here's how it could work.

这曾经是我的独特卖点，但现在消失了。

And that's that was kind of my my unique selling point, and that's gone.

任何人都能做我以前能做的事。

Anyone can do what I could do.

你知道吧？

You know?

但你仍然需要学会在什么时候适合做原型，如何思考原型设计，以及如何获取工具来构建有用的原型，以便用来探索各种可能性。

It's like but but but it does but you still have to learn when it's appropriate to prototype, how to think about prototyping, how to get the tools to build useful prototypes that you can you can use to explore things.

我非常兴奋地向大家介绍本季的赞助商——Vanta。

I am so excited to tell you about this season's supporting sponsor, Vanta.

Vanta 帮助包括 Cursor、Ramp、Duolingo、Snowflake 和 Atlassian 在内的超过 15,000 家公司赢得并证明客户的信任。

Vanta helps over 15,000 companies like Cursor, Ramp, Duolingo, Snowflake, and Atlassian earn and prove trust with their customers.

由于人工智能的发展，团队正在以前所未有的速度构建和发布产品，但随之而来的，是产品和业务中引入的风险比以往任何时候都更高。

Teams are building and shipping products faster than ever thanks to AI, but as a result, the amount of risk being introduced into your product and your business is higher than it's ever been.

我接触过的每一位安全负责人，都深感保护公司、业务，更不用说客户数据的压力日益加重。

Every security leader that I talk to is feeling the increasing weight of protecting their organization, their business, and not to mention their customer data.

由于一切进展太快，他们不得不不断应对突发状况，猜测优先级，并勉强使用过时的解决方案。

Because things are moving so fast, they are constantly reacting, having to guess at priorities, and having to make do with outdated solutions.

Vanta 通过自动化合规与风险管理，支持超过 35 个安全与隐私框架，包括 SOC 2、ISO 27001 和 HIPAA。

Vanta automates compliance and risk management with over 35 and privacy frameworks, including SOC two, ISO 27,001, and HIPAA.

这帮助公司快速实现合规并持续保持合规。

This helps companies get compliant fast and stay compliant.

与以往任何时候相比，信任更能决定你业务的成败。

More than ever before, trust has the power to make or break your business.

了解更多，请访问 vanta.com/lenny。

Learn more at vanta.com/lenny.

作为本播客的听众，您可以享受 Vanta 1000 美元的优惠。

As a listener of this podcast, you get $1,000 off Vanta.

那就是 vanta.com/lenny。

That's vanta.com/lenny.

我要岔开一下话题。

I'm gonna take a tangent.

你的 AI 技术栈中包含哪些工具？

What's what's kind of in your stack, your AI stack?

你最常使用哪些模型？

What models are you using most?

你觉得哪些工具有用？

What tools do you find useful?

目前我主要用Claude。

So right now, I'm mostly Claude.

我大量使用Claude进行编码工作。

I do a huge amount of work using Claude code.

我主要还是个Claude编码用户，但Claude编码有两个方面是我常用的。

Well, I'm I'm mainly still a Claude code person, but there are two sides of Claude code that I use.

一个是运行在你电脑上的Claude编码，另一个是网页版的Claude编码，也就是他们托管的版本。

There's the Claude code that runs on your computer, And then there's Claude code for web, which is their hosted version of Claude code.

我更常使用网页版，部分原因是它可以通过手机访问。

And I use that one more than the one on my own computer, partly because that's the one you can access through your phone.

如果你在iPhone上安装了Anthropic Claude应用，里面有个编码标签页，你可以进去让它帮你写代码。

If you've got the Anthropic Claude app installed on iPhone, there's a code tab, and you can go in there, and you can tell it to write you things.

而它是在他们的服务器上运行的。

And that it's running on their servers.

你需要提供一个你自己的 GitHub 仓库，让它能在其中工作。

You get need to give it a GitHub repository of yours that it can work within.

但从安全角度来看，这也很棒，因为如果你在笔记本电脑上运行 Claude 代码，可能会发生一些坏事。

But it's also great from a security point of view because if you're running Claude code on your laptop, there's risks that bad things can happen.

它可能会不小心删除一些东西。

It might accidentally delete things.

如果我在 Anthropic 的服务器上运行，我根本不在乎。

If I'm running on Anthropic servers, I couldn't care less.

毕竟，那是他们的电脑。

Like, it's their computer.

不是我的电脑。

It's not my computer.

随便用吧。

Go wild.

这意味着你可以以随心所欲的方式运行这些操作。

So this means that you can run these things in, in YOLO mode.

这被称为危险地跳过权限，Claude就是这样说的。

This is, Claude calls it dangerously skip permissions.

OpenAI 确实称之为 YOLO。

OpenAI actually do call it YOLO.

他们提供了这个选项。

They've got an option for that.

在这种模式下，智能体不会每次都问你是否要执行某个操作。

And that's the mode where the agent doesn't ask you if it should do something all the time.

这是一款完全不同的产品。

And that is a different product.

我认为，很多还没有接触编码智能体的人，还没试过在不安全模式下使用它们。

I think a lot of people who haven't got on board with coding agents yet haven't tried them in the unsafe mode.

他们使用的编码智能体总是像在问：‘我可以运行这段代码吗？’

They're using coding agent where it's like, oh, can I run this piece of code?

我可以编辑这个文件吗？

Can I edit this file?

这意味着你必须全程全神贯注地盯着它。

And that means you have to pay complete attention to it the whole time.

这就像是在和一个特别烦人的幼儿打交道，它不停地缠着你问想做什么。

And it's like working with a really frustrating toddler that's constantly nagging you about what it wants to do.

一旦解除安全限制，我就能同时运行四个，然后去喝杯茶，回来时它们已经为我完成了些有用的事情。

The moment you take the safeties off, now I can run four of them and go and have, like, go and go and have a cup of tea and come back and they've they've achieved something useful for me.

但这种方式本质上是不安全的。

But it's inherently unsafe.

如果它在Claude Code for Web中运行，最坏的情况可能只是不小心泄露了你的私有源代码。

If it's running in Claude Code for Web, the only bad thing that could happen is maybe it accidentally leaks your private source code.

而我的代码都是开源的，所以我并不在意。

And my code is all open source, so I don't care.

这其实是个挺有用的技巧。

That's that's a useful trick there.

但确实如此。

But yeah.

所以我会在手机上使用它。

So I use that on my phone.

我经常同时运行两三个这样的程序。

I often have two or three of those running.

我的许多主要项目都是通过在手机上提示完成的。

A lot of my major projects are done mostly prompting on my phone.

如果涉及安全或非常重要，我可能会把它移到笔记本电脑上进行后续详细审查。

If it's security adjacent or super important, I might pull it down to my laptop to do a thorough review later on.

但大多数审查工作都可以通过 GitHub 完成。

But most of the review you can do through GitHub.

这些工具会提交拉取请求，然后你可以使用审查他人代码的相同工具来审查智能体生成的代码。

Like these things will file pull requests, and then you use the same tools you'd use to review code from other people to review the code from the agents.

话说回来，OpenAI 大约三周前发布了 GPT 5.4。

That said, OpenAI came out with GPT 5.4 about three weeks ago.

它非常、非常、非常好。

It's very, very, very good.

我认为它与Cloud Opus 4.6不相上下，甚至可能更好。

I think it's on par with Cloud Opus 4.6 and possibly even better.

这些公司一直在不断超越彼此。

These companies are constantly leapfrogging each other.

所以我最近更多地依赖它，而且也更便宜。

So I have been use leaning back it's also cheaper.

所以这个月我更多地使用了GPT 5.4和OpenAI Codex。

So I've been leaning on GPT 5.4 a lot more this month and OpenAI Codex.

现在OpenAI Codex和CloudCode几乎完全难以区分。

And OpenAI Codex and CloudCode are almost almost indistinguishable from each other now.

它们都是非常非常优秀的软件。

They're both very, very good pieces of software.

我对此早有预料。

And I kind of expect this to happen.

比如，下一个Gemini模型发布后，可能会在几个月内成为最好的编程模型，到时我可能会转用那个生态系统。

Like, the next Gemini model comes out might be become the best coding model for a couple of months, in which case I might switch myself into that ecosystem.

部分原因是我自己也会写这方面的内容。

Partly because I write about this stuff as well.

我喜欢尽可能熟悉各种产品。

I like to stay familiar with as many of the offerings as possible.

但我还是不断回到Claude Code，主要是因为它符合我的品味。

But I keep on coming back to Claude Code mainly because it fits my taste.

有个挺奇怪的现象：我对代码的运作方式有非常特定的偏好，而巧合的是，这恰好与Claude Code的风格一致，这还挺有意思的。

Like, there's this weird thing where I've got a very specific taste in how I like code to work, which coincidentally happens to map to how Claude code likes to work, which is kind of interesting.

GPT 5.4几乎也符合我的口味，但还差那么一点。

And g p d 5.4, it's almost matches my taste, but not quite.

也许这是因为我和Claude相处的时间更长。

And maybe that's because I've just spent more time with Claude.

我的提示方式已经逐渐更适应Claude的思维方式。

My prompting style has evolved more to fit the Claude way of thinking.

我不确定。

I don't know.

这些东西也很奇怪。

This stuff's also weird.

一切都是感觉驱动的。

It's vibes all the way down.

这太有趣了。

That is so interesting.

所以你指的是代码的风格，它生成的代码质量，而不是对话方式或用户体验。

So the taste is the code, the quality of the code it puts out is what you're talking about, not like the conversation and the UX.

当然。

Absolutely.

我不在乎它们怎么跟我对话。

Don't care about how they talk to me.

我只是用它们来完成工作。

Like I'm using them to get stuff done.

是的。

Yeah.

是的。

Yeah.

因为我在听你说话的时候就在想，是什么会让一个人持续使用某个模型？

Because I was thinking as you're talking, what is the thing that will get someone to stick with a model?

可能就是你所描述的代码质量，它写代码的方式。

And it could be what you're describing the quality, the way it writes code.

也可能是用户体验，或者是对话的氛围。

It could be the UX, it could be the conversation vibes.

嗯，

Well,

最能让人粘住的东西应该是记忆功能。

the stickiest thing is meant to be memory.

比如，它们都有这些功能，会记住关于你的事情，但我非常讨厌这些功能，只要有可能我就会关掉，主要是因为作为一名AI研究员，我需要在提问时看到其他人看到的相同内容。

Like, the the all of the they they all have these features where they will remember things about you, and and I hate those features, and I turn them off wherever I can because mainly because as an AI researcher, I need to see what everyone else sees when I'm prompting.

我不想对全世界说：天啊，看，这个东西现在能用了。

Like, I don't want to say to the world, oh my goodness, look, this thing works now.

这对我有效，是因为它基于我之前的一些对话内容。

And it only works for me because it's based on previous, like, into previous conversations that I've had.

也许我因此错过了一些非常重要的东西。

And maybe I'm missing out on something really important there.

但记忆功能正是所有实验室都在努力用来增强用户粘性的那个功能。

But the the memory feature is is is that thing that all of the labs are trying to be more sticky with.

话说回来，几周前当OpenAI的军事相关事件发生时，Anthropic趁机宣传说：嘿，为什么不转用Claude呢？

That said, when the whole the the OpenAI military stuff happened a few weeks ago, Anthropic track took advantage by saying, hey, why don't you move to Claude?

他们的方式是在Claude的入门页面上写道：点击这个按钮，将你的记忆从ChatGPT转移过来，然后粘贴到ChatGPT中。

And the way they did that is they had a Claude onboarding page that said, transfer your memories from ChatGPT by clicking this button and then pasting it into ChatGPT.

但那只是一个提示语。

And it was just a prompt.

他们设置了一个提示语，内容是：嘿，ChatGPT。

They had a prompt, which was, hey, chat g p t.

告诉我你记得的关于我的所有事情。

Tell me everything that you've remembered about me.